Assessing the extent, stability, purity and properties of silanised detonation nanodiamond

Applied Surface Science 357 (2015) 397–406

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Assessing the extent, stability, purity and properties of silanised detonation nanodiamond Emer Duffy a , Dimitar P. Mitev a , Stuart C. Thickett a , Ashley T. Townsend b , Brett Paull a,c , Pavel N. Nesterenko a,∗ a

Australian Centre for Research on Separation Science, School of Physical Sciences, University of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia Central Science Laboratory, University of Tasmania, Private Bag 74, Hobart, Tasmania 7001, Australia c ARC Centre of Excellence for Electromaterials Science, School of Physical Sciences, University of Tasmania, Hobart, Tasmania 7001, Australia b

a r t i c l e

i n f o

Article history: Received 13 March 2015 Received in revised form 26 August 2015 Accepted 1 September 2015 Available online 2 September 2015 Keywords: Detonation nanodiamond Surface modification Silanisation Dispersion stability Elemental composition Thermal stability

a b s t r a c t The functionalisation of nanodiamond is a key step in furthering its application in areas such as surface coatings, drug delivery, bio imaging and other biomedical avenues. Accordingly, analytical methods for the detailed characterisation of functionalised nano-material are of great importance. This work presents an alternative approach for the elemental analysis of zero-dimensional nanocarbons, specifically detonation nanodiamond (DND) following purification and functionalisation procedures. There is a particular emphasis on the presence of silicon, both for the purified DND and after its functionalisation with silanes. Five different silylation procedures for purified DND were explored and assessed quantitatively using inductively coupled plasma-mass spectrometry (ICP-MS) for analysis of dilute suspensions. A maximum Si loading of 29,300 ␮g g−1 on the DND was achieved through a combination of silylating reagents. The presence of 28 other elements in the DND materials was also quantified by ICP-MS. The characterisation of Si-bond formation was supported by FTIR and XPS evaluation of relevant functional groups. The thermal stability of the silylated DND was examined by thermogravimetric analysis. Improved particle size distribution and dispersion stability resulted from the silylation procedure, as confirmed by dynamic light scattering and capillary zone electrophoresis. © 2015 Elsevier B.V. All rights reserved.

1. Introduction There is a growing interest in the properties and applications of ultra-dispersed matter, specifically nanodiamonds (ND), which are produced by a variety of synthetic methods. Some popular production methods include cavitation and laser production, chemical vapour deposition, crushed high-pressure high-temperature (HPHT) diamond, and detonation synthesis for the production of detonation nanodiamond (DND) [1]. The method of production can have a significant impact on the resultant diamond’s properties, and DND is known to have higher impurity levels than those

Abbreviations: BET, Brunaer–Emmett–Teller; BGE, background electrolyte; CE, capillary electrophoresis; CZE, capillary zone electrophoresis; DMODCS, dimethyloctadecylchlorosilane; DND, detonation nanodiamond; EOF, electroosmotic flow; HMDS, hexamethyldisilazane; ICP-MS, inductively coupled plasma mass spectrometry; TGA, thermogravimetric analysis; TMCS, trimethylchlorosilane. ∗ Corresponding author at: Australian Centre for Research on Separation Science (ACROSS), School of Chemistry, Faculty of Science Engineering and Technology, University of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia. E-mail address: [email protected] (P.N. Nesterenko). http://dx.doi.org/10.1016/j.apsusc.2015.09.002 0169-4332/© 2015 Elsevier B.V. All rights reserved.

present in other synthetically derived diamonds [2]. The type and level of impurities present in DND can vary quite dramatically, and they depend on both the production and purification methods utilised [3]. DND has shown significant promise for application in a wide variety of areas, including materials synthesis within polymer and carbon-based nano-composites [4–6], coating technologies [7], separation science/chromatography [8], and biomedical applications such as drug delivery and bio-imaging [9–12]. The purity of DND is of the utmost importance in biomedical applications in terms of furthering research outputs and reaching the ultimate goal of using these particles in human patients. It is therefore vital that reliable methods of elemental analysis and evaluation for DND are available. Furthermore, the control of DND surface properties and functionalities will be a key step in ensuring their continued progress in real world applications. It is imperative that there is full understanding and control over the surface chemistry of DND to allow its dispersion in solvents, or bonding to solid matrices, while simultaneously preventing the problem of agglomeration. Silylation can be very useful as a tool to tailor diamond nanoparticles for use as composite ingredients [13–15], or to adjust the surface

398

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

properties of prospective diamond-based sorbents for chromatography [8,16,17]. It offers a method for surface homogenisation prior to bio-conjugation. It is important to note the de-agglomeration of primary DND particles is highly important before proceeding with surface modification, and silylation has proven to be an effective route to improving dispersion stability and reducing aggregate size as demonstrated by Krueger et al. where the silylation of nanodiamond with (3-aminopropyl)trimethoxysilane and further with biotin was carried out for bioapplications such as labelling and drug delivery [18], and with 3-acryloxypropyl(trimethoxysilane) in combination with beads-assisted sonication for improving dispersion quality and stability [19]. Jeong et al. also reported on the silylation of ND for improving dispersion stability, prior to its inclusion in development of transparent composites where aggregate size directly affected the composite material’s performance [20]. Dolmatov described a silylation procedure for nanodiamonds aimed at composites based on elastomer and polymer matrices, however there was no characterisation of the actual degree of silylation [21]. In the preparation of such stabilised, homogenised or modified particles, it is also vital that there are reliable characterisation methods allowing an understanding of the surface chemistry and impurities present in DND, as well as evaluating the stability of such particles under different conditions. Until as recently as 2013, there was a lack of comprehensive information on the presence of metal and non-carbon impurities in DND as a whole. Controlling and quantifying the heteroatom content in DND is a key step towards fully understanding and ensuring a high degree of purity before applying this variable material. There was an absence of suitable reliable analytical methods for this specific nanocarbon; an issue that was recently addressed in our laboratories. The direct aspiration of very dilute DND suspensions into a sector field inductively coupled plasma mass spectrometer (ICP-MS) has proven to be a reliable and sensitive method for the determination of impurities in DND [22,23]. Herein we demonstrate the application of this method to the quantification of purposely introduced elements (like Si) to the surface of DND. This analysis was performed on silylated DND, Si-contaminated DND and DND that underwent different deep purification routes in order to clearly assess the effect of silylation on the DND. Although requiring expensive instrumentation and experienced operators, the method is relatively fast and reduces the reliance on more time-consuming and less informative methods such as thermogravimetric analysis (TGA), FTIR, and others [24–26], which require further analysis to allow quantification of heteroelements in DND. In this instance, TGA and FTIR were used to further characterise the surface chemistry and thermal stability of the modified DND. In addition, the stability of modified DND was investigated using capillary zone electrophoresis (CZE), which has proven to be an informative technique with regard to nanoparticle size distributions, dispersion quality and stability or propensity to form aggregates, which is reflected in the peak shape obtained by a voltage-induced separation [27]. 2. Experimental 2.1. Materials and reagents The detonation nanodiamond (DND) was obtained from YTM ARGE A.S. (Istanbul, Turkey; product code NDG.11.02.28.01). Additionally, raw detonation soot (DS) was supplied by the same producer; in the form of a black-coloured powder. A 1% aqueous suspension of the DS had a pH of 9.43. Milli-Q deionised water (Millipore, Bedford, MA, USA) with resistivity 18.5 M cm (at 298 K) was used throughout this work. All reagents used were of analytical grade quality: 95–97% H2 SO4 and 48% HF, 70% HNO3 were all obtained from Chem-Supply,

Gillman, SA, Australia. In addition, 37% HCl from Merck, Kilsyth, Australia, and 70% HClO4 from Univar, Ingleburn, Australia were used. The silylation reagents: trimethylchlorosilane (TMCS), dimethyloctadecylchlorosilane (DMODCS), hexamethyldisilazane (HMDS), and toluene were of analytical grade quality, and were purchased from Sigma Aldrich, Castle Hill, NSW, Australia. Sodium tetraborate (99.998%) and NaOH were used in the preparation of background electrolytes (BGE) for CZE, and were also obtained from Sigma Aldrich, Castle Hill, NSW, Australia. Buffers were prepared by dilution of a stock solution of sodium tetraborate, and the pH was adjusted using 1 M NaOH. 2.2. Modification of detonation nanodiamond 2.2.1. Initial purification procedures The raw DS obtained from YTM ARGE A.S. had been purified by the supplier prior to shipment, whereby a sulphuric acid/potassium dichromate wet graphite digestion method was utilised [19]. This material was additionally purified on-site using a procedure previously described by our research group [28] for preparation of NSFPA type of purified DND. Briefly, NSFPA was purified from the DS using acidic oxidation with a mixture of HNO3 , H2 SO4 , HF and HClO4 (where % ratio was 44:44:6:6). Subsequent washing steps were performed to reduce the digested impurities and residual acids. It is important to note that Si contamination of DND can occur through everyday processing of DND materials in borosilicate glassware [28] and as a precaution, special attention was paid to handling procedures for DND within this study wherein only polypropylene containers were used for storage of DND samples. Obviously, contamination does not occur in glassware when nonpolar organic solvents like toluene are used as the modification medium, as described below for modification of DND. 2.2.2. Hydrophobisation of detonation nanodiamond using silanisation The NSFPA type of purified DND was used in all experiments on hydrophobisation. Sample 1. 0.451 g of DND, and 5 mL TMCS were added to 45 mL of as-stored toluene in a triple-neck flat-bottomed flask. The reaction was kept under constant magnetic stirring and the reaction flask was fitted with a reflux condenser and a thermometer. The temperature was set to 376 K, at which time an additional 1 mL of TMCS was added to the flask. The mixture was then stirred at 376 K for 21 h. The subsequent additions of 2 mL of TMCS and 2 mL pyridine were made, and the reaction was allowed to continue for a further 5 h. Sample 2. DND was firstly dried under vacuum at 673 K overnight. Then 0.218 g of dried DND and 0.207 g DMODCS were added to 50 mL of dehydrated toluene in a triple-neck flat-bottomed flask. The reaction flask was fitted with a reflux condenser and a thermometer, as above, and was kept under constant magnetic stirring. The temperature was then set to 373 K, at which time an additional 0.313 g DMODCS dissolved in 5 mL of toluene was added. The reaction mixture was refluxed under stirring for 19 h. Sample 3. 0.264 g of dried DND (vacuum dried, 673 K overnight), 4 mL of HMDS and 8 mL of TMCS were added to 188 mL of dehydrated toluene in a triple-neck flat-bottomed flask. The reaction mixture was refluxed under stirring for 19 h. Sample 4. 0.189 g of dried DND (vacuum dried, 673 K overnight), 0.534 g DMODCS and 2.2 mL of HMDS were added to 58 mL dehydrated toluene in a triple-neck flat-bottomed flask. The reaction mixture was held under the same conditions as described above for sample 3. Sample 5. 0.220 g of dried DND (vacuum dried, 673 K overnight) and 10 mL TMCS were added to 50 mL dehydrated toluene in a

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

399

triple-neck flat-bottomed flask. The reaction mixture was held under the same conditions as described above for sample 3.

Ti

V

Al K

10000

W

Na

1000

2.3. Instrumentation

Mo 100

An Element 2 sector field ICP-MS spectrometer (Thermo Fisher, Bremen, Germany) was employed for ICP-MS analysis of DND. Aqueous suspensions of DND were diluted to 0.1 mg g−1 prior to direct analysis by ICP-MS according to the procedure described in [22]. This approach was applied for multi-element detection, and results are reported for elements that were present in DND at concentrations higher than 10−5 mass% in the dry DND, or 10−8 mass% in suspension. Multiple spectral resolutions were employed to allow commonly encountered polyatomic interferences to be clearly resolved from analytes of interest. We note that non-metal impurities like O, N and Cl are excluded from consideration here, as they are not determined by ICP-MS. Surface area measurements were performed using a surface area analyser (Tristar II 3020, Micromeritics, Gemini, GA, USA) by the nitrogen adsorption/desorption method. FTIR analysis was performed using a Spectrum 100 FT-IR Spectrometer fitted with Universal ATR Accessory (PerkinElmer, Waltham Massachusetts, USA). The scanning range was 650–4000 cm−1 and the resolution was 1 cm−1 . X-ray photoelectron spectroscopy. An ESCALAB 250Xi XPS (Thermo Scientific, UK) incorporating a 165 mm hemispherical electron energy analyser was used. The incident radiation was monochromatic A1 X-rays (1486.6 eV) at 150 W (13 kV, 12 mA). Survey (wide) scans were taken at an analyser pass energy of 100 eV and (narrow) higher resolution scans at 20 eV. Survey scans were carried out over 1360–0 eV binding energy range with 1.0 eV steps and a dwell time of 100 ms. Narrow higher resolution scans were run with 0.1 eV steps and 250 ms dwell time. Base pressure in the analysis chamber was below 1.5 × 10−9 Torr and during sample analysis 1.0 × 10−8 Torr. The data were analysed using the Avantage software suite (Thermo Scientific, UK). A Zetasizer Nano ZS particle analyser (ATA Scientific, Taren Point, NSW, Australia) fitted with 632.8 nm red laser and 175◦ backscatter detector was used to estimate zeta potential and particle size distributions of derived samples. DND suspensions in isopropanol and water were analysed after 1-min of sonication in an ultrasonic bath (Bransonic 5510E-DTH, Branson, Danbury, USA). TGA data were obtained with a Labsys Evo instrument (Setaram, Caluire, France) under nitrogen flow (50 mL min−1 ). DND samples were placed in alumina crucibles for analysis, where a heating rate of 2.5 ◦ C min−1 was applied from 30 to 900 ◦ C. CZE experiments were performed on an Agilent 7100 CE system (Agilent Technologies, Germany) equipped with a photodiode array detector. Adsorption was monitored at 230 nm at a data acquisition rate of 40 Hz. All reported results were obtained in triplicate. Fused silica capillaries (Polymicro, Phoenix, AZ, USA) with 75 ␮m i.d., 365 ␮m o.d. were used, with a total capillary length of 32.5 cm. Detection windows were burned 8.5 cm from the capillary end using a butane torch. Pre-conditioning of capillaries was performed using 1 M NaOH (15 min flush), followed by H2 O (10 min) and finally with BGE (10 min). DND suspensions of 0.125 mg mL−1 were injected for 5 s at 5 kPa, and separated under an applied potential of +5 kV. 3. Results and discussion 3.1. DND purification and refinement It is important to consider the possible presence of impurities in DND materials, prior to modification of their surface. Ensuring an effective purification route is followed, and good dispersion

Mn

B

10

Si 1

Fe

0.1 10000 1000

100

10

Ca

1

Ni 0.1

1

10

100

1000 10000

1

S 10

Mg

Cr

100

Sn 1000

Zr Cu NSFPA/YTM YTM

10000

Zn

Pb

Sb Ba

Fig. 1. Impurity contents (␮g g−1 , logarithmic scale) in the samples of commercial DND as received from YTM ARGE A.S. and after purification using NSFPA protocol (initial purification procedure as described in Experimental methods) [28].

stability is achieved with small particle sizes (no agglomerates) is a key step in furthering the functionalisation of DND for applications in biomedicine and composite materials. The presence of Si and other non-carbon impurities were assessed by ICP-MS, with the quantitative findings shown in Fig. 1. As a consequence of the dichromate/acid-based purification that had been carried out on the YTM material prior to receiving it, this DND product was abundantly rich in impurities with an Si-level of 2870 ␮g g−1 for the sample in suspension (Fig. 1). Prior to the controlled silylation of DND particles, a refinement procedure was applied to the materials to ensure contributions from Si impurities were minimised. Conventional handling of DND materials during production procedures and everyday laboratory work can actually significantly increase their Si-content [28]. Following the refinement procedure, the Si content was reduced to only 2.2 ␮g g−1 (Fig. 1). The Brunauer–Emmett–Teller (BET) surface areas for DND materials were calculated from the nitrogen adsorption isotherms and were found to be 260 ± 5 m2 g−1 for YTM and following refinement, a higher surface area of 287 ± 8 m2 g−1 was obtained for the NSFPA treated DND as expected due to the additional purification procedure resulting in further removal of surface impurities, and a reduction in contamination levels up to a factor of 6.5 [28]. Elevated Al and S contents for the refined DND were suggested to result from contaminated acid. 3.2. Silylation Surface hydroxyl and carboxyl groups of DND may be successfully derivatised following a silylation attack, as is similarly performed on a range of chemical compounds, glass, silica and other surfaces and molecules. The application of the ICP-MS characterisation method [22] permitted an accurate measurement of the silylation degree on the DND surface, and thus provided valuable quantitative information for comparison of the effectiveness of the different silylating agents, and their corresponding effects on DND impurity levels. Silylation is highly sensitive to the presence of moisture in the reaction mixture. DND is a hygroscopic material (dry DND kept in air can retain up to 4.1% of water) [29], so all samples were dried under vacuum at 400 ◦ C overnight prior to silylation. As a demonstration, the “as-stored” DND and toluene were not subjected to

400

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

A

Ti

V

B

Al

W

Na

K Na

1000

Mo

Mo 100

Mn

Al

10000

W

1000

Ti

V

K

10000

100

B Mn

10

B

10

Si

Si 1

Fe

0.1

0.1 10000 1000

100

10

1

Ca

0.1

1

10

100

10000 1000

1000 10000

100

10

Ca

1

100

1000 10000

S Cr Zr

Zr

10000

Cu

10000

Sb Pb

10

1000

1000

Zn

1

100

Cr Sn

Cu

0.1

10

Mg

100

Sn

1 1

S

10

Mg

1

Fe

Zn

Ba

NSFPA untreated NSFPA treated with TMCS + HMDSz NSFPA treated with TMCS

Sb Pb

Ba

NSFPA treated with DMODCS NSFPA treated with mixture of DMODCS+HMDSz NSFPA untreated

Fig. 2. Composition of minor elements (␮g g−1 , logarithmic scale) of original DND type NSFPA and the silylated products.

Table 1 Elemental concentrationsa in dry DND samples by ICP-MS, ␮g g−1 . Procedure Elementb

– NSFPA

– YTM powder

(1) TMCS

(5) TMCS

(2) DMODCS

(4) DMODCS + HMDS

(3) TMCS + HMDS

B Sr Zr Mo Sn Sb Ba W Pb Bi Na Mg Al Si P S Ca Sc Ti V Cr Mn Fe Cod Ni Cu Zn K As Total

31.1 0.1 0.2 1.2 0.0 19.3 234 0.3 96.5 0.0 6.6 0.0 0.0 48.0 0.0 189 45.5 0.0 429 5.0 1.0 0.2 35.0 0.0 0.0 0.7 0.3 0.0 0.0 1142

6.4 9.1 0.5 44.3 36.9 0.4 49.9 4.0 11.3 0.1 244 264 460 1430 3.5 764 3020 0.0 1280 27.7 5110 49.3 328 0.3 3.4 74.9 61.9 207 0.0 13,480

36.8 0.3 0.3 1.9 1.4 28.2 238 0.7 81.9 0.0 30.0 6.0 8.3 212 0.4 254 64.4 0.0 571 7.5 1.5 0.6 41.8 0.1 0.2 1.9 32.1 0.0 0.0 1621

32.8 1.0 0.3 0.9 1.5 19.0 229 0.1 84.5 0.0 21.5 10.9 4.8 7020 5.1 244 942c 0.0 605 8.0 1.7 1.2 64.9 277 0.0 5.2 14.0 3.7 0.0 9593

28.1 0.2 0.2 1.1 2.6 22.7 214 0.3 76.3 0.0 5.8 2.7 0.0 2770 0.0 180 54.9 0.0 488 6.1 8.5 0.5 41.9 655 0.8 1.5 5.7 0.0 0.0 4565

25.6 0.2 0.4 1.4 0.7 22.5 208 0.1 59.3 0.0 19.8 1.1 0.4 27,400 3.0 208 69.7 0.0 524 7.5 1.7 0.8 60.9 950 1.4 2.0 8.3 2.3 0.0 29,560

35.3 0.2 0.5 1.3 3.6 24.0 206 0.2 49.1 0.0 13.6 1.7 0.0 29,230 0.7 168 30 0.0 520 7.9 1.9 0.8 54.6 888 0.9 1.8 6.4 0.0 0.0 31,312.1

a b c d

Excluding O, N, Cl, etc. not determined by ICP-MS. In the order of instrument measurement. Contamination of the sample. Elevated concentration of Co in silylated DND originates from the silica gel used for drying of toluene, namely the “blue → red” moisture indicator).

any drying steps (procedure outlined for sample 1 above), resulting in the reaction being blocked, despite the presence of pyridine base which acts as a HCl acceptor and theoretically facilitates the silylation [30]. ICP-MS analysis of silylated DND allowed quantification of Si and other minor elements. The results shown in

Fig. 2 highlight that without an initial drying step, sample 1 (TMCS treated) only showed a 5-fold increase in Si content to 212 ␮g g−1 from the 48 ␮g g−1 originally found in the starting DND (Table 1). There was a significant increase in the Si content to 7020 ␮g g−1 recorded for sample 5, which was dried under vacuum at 400 ◦ C

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

1200

Fitrate

401

Total suspension of DND in water

Si content, ng g-1

1000

800 S

600 S 400 S 200

S

0 1

0

6

24

360

Time, hours

0

1

6

24

360

Time, hours

Fig. 3. Kinetics of contamination of NSFPA by Si during storage (of aqueous suspension of DND (0.66 mg g−1 ) and filtrate of the same DND suspension) in borosilicate laboratory glassware (S means sonication). Adapted from [28].

overnight and silylated at higher concentration of TMCS in toluene dried over silica gel (see Fig. 2B). This was achieved in the absence of organic bases, which are normally used to accelerate the reaction by absorbing released HCl as shown in Eqs. (1) and (2). Other elemental impurities also present in the DND after each treatment were determined and are shown in Table 1. The addition of HMDS to the reaction (sample 3) enhanced the silylation level of DND. This effect is connected with the release of ammonia as a product of reaction (Eq. (4)), which neutralises HCl released from reaction of TMCS and protogenic groups at the surface of DND (Eqs. (1) and (2)). The Si-content increased dramatically to 29,300 ␮g g−1 in sample 4, realizing an increase of over 610 times as compared to the original DND. C OH + ClSi(CH3 )3 →

C O Si(CH3 )3 + HCl ↑,

(1)

C C( O)OH + ClSi(CH3 )3 →

C( O) O Si(CH3 )3 + HCl ↑,

(2)

C NHR + ClSi(CH3 )3 →

+

(3)

−

C N HRSi(CH3 )3 Cl ,

Fig. 4. ATR-FTIR spectra of NSFPA and five silylated samples.

2 C OH + (CH3 )3 SiNHSi(CH3 )3 → 2 C O Si(CH3 )3 + NH3 ↑, (4) where C OH, C C( O)OH and C NHR are protogenic functional groups at the surface of DND. A similar result was obtained when using the bulkier derivatisation agent DMODCS instead of TMCS. When applied as the sole silylating agent upon derivatisation of vacuumed DND in dry toluene, a significant increase in the Si-content to a level of 2770 ␮g g−1 was observed in sample 2, which is 58 times higher than in NSFPA type DND. However, when DMODCS was combined with HMDS, it resulted in a further increase in Si content up to 27,400 ␮g g−1 , i.e. 570 times, as was similarly observed for the combination of silylating reagents discussed above for sample 3. The maximum DND surface coating of 0.88–0.95 mmol g−1 as calculated from the increase in Si-content after silylation is 2–3 times higher than the coating of 0.3–0.4 mmol g−1 reported for the hydrophobic DND samples prepared by esterification with alkanoic acid chlorides [31]. It should be noted that a serious contamination of DND by Si is possible during everyday processing of DND in glassware [28]. Fig. 3 shows the kinetic of accumulation of Si during storage of aqueous suspension of DND in deionised water in borosilicate

glassware (for both the DND suspension, and the water after filtration of the suspension) due to exposure to the glass. The maximum contamination could be as high as 2–3 mg g−1 as compared with 47 ␮g g−1 detected in NSFPA type DND. Special precautions were taken during modification processes of DND within this work as described in the purification and modification procedures. 3.3. Infrared spectroscopy of silylated DNDs FTIR was used to confirm derivatisation of the DND surface. Fig. 4 shows several characteristic absorbance bands in the spectra of the purified and silylated DNDs. The line at 848 cm−1 is associated with rocking vibrations of CH2 -groups [32], and can also be linked directly with the Si CH3 moiety [33]. There was a gradual increase of hydrocarbon groups as the extent of silylation increased (samples 5 → 4 → 3) as confirmed by these vibrations, and by the C H asymmetric stretch around 2960 cm−1 and stretches corresponding to Si CH3 bands in Si(CH3 )2 O groups within the silane structure at 1260 cm−1 . These observations support the increasing extent of silylation previously discussed in relation to ICP-MS characterisation. The broad spectral line at 1435 cm−1 is persistent

402

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

Fig. 5. XPS spectra for purified nanodiamond (NSFPA) and silylated nanodiamond (Sample 4). The C1s, O1s and Si2p core level spectra are shown plotted above the related fitting components.

for DND samples treated by the HMDS reagent, and according to Nascimento Filho et al. [34] this spectral line can be attributed to physically adsorbed HMDS. Adsorption at 1707 cm−1 is characteristic for all silylated nanodiamond samples and may be attributed to the (C O) band of trimethylsilyl-esters formed with carboxylic groups at the diamond surface [33]. It should be noted that carboxylic groups are more acidic and reactive as compared with hydroxyl-groups and residual water, so this band was noted for all samples independent of their respective modification conditions. The peak shift from 1110 cm−1 (NSFPA) to 1092 cm−1 for the silyl derivatives confirms the formation of Si O bonds [35].

3.4. X-ray photoelectron spectroscopy X-ray photoelectron spectroscopy (XPS) was also performed on purified NSFPA-type DNDs as well as after silylation (sample 4). XPS analysis enables a more detailed understanding of surface functionalisation as well as carbon hybridisation at the DND interface, given that it is well known that a graphitic layer consisting of sp2 carbon forms on the nanodiamond surface [36–38]. Furthermore, given the average radius of the DNDs in this work, we are able to probe the majority of the DND nanoparticle volume by this technique [39]. The XPS survey scan of purified DNDs (NSFPA, Fig. 5) consisted of a dominant C 1s peak (89.56 atomic %) and smaller contributions from O 1s (8.75%), N 1s (1.43%) and Cl 2p (0.26%); the N and Cl contributions are attributed to the synthetic method and purification of the nanodiamonds.

High-resolution spectra of the C 1s and O 1s region (Fig. 5) provide significant information as to the surface composition of the purified DNDs. De-convolution of the C 1s spectra is well fitted by two peaks centred at 284.8 eV and 285.7 eV, respectively (both with a FWHM of 1.9 eV), which are attributed to an sp2 -hybridised graphitic shell and an sp3 -hybridised diamond core, respectively. (These peak maxima differ slightly to those reported by Butenko et al. [40] and Petit et al. [38], however a high-quality peak fit was not possible using reported values). The fractional peak area attributed to sp2 -hybridised carbon is ∼55%, which is particularly high compared to reported XPS analyses of other DND samples [36,38]. The O 1s spectra is also fitted by two main populations at 529.7 eV and 531.5 eV, which are likely due to unsaturated C O groups and saturated C O groups (86% peak area) at the DND surface. XPS analysis thus suggests that hydroxylic or phenolic groups are the dominant oxygen-containing species present in the purified DNDs. Co-silanisation with HMDS and DMODCS (sample 4) reveals important structural changes to the samples as analysed by XPS (shown in Fig. 5). Si is detected at 0.48 atomic %, in addition to a 0.64% decrease in the O 1s signal (with both oxygen peaks decreasing to an equivalent extent). There is also a significant decrease in the C 1s peak area attributed to sp2 hybridised carbon (down to 40%), despite the total carbon content remaining approximately constant, which suggests effective nanodiamond functionalisation, given that the silanising agents also introduce sp3 hybridised carbon atoms to the particle surface. The high resolution Si 2p spectrum for sample 4 also indicates two Si populations,

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

403

tendency for aggregation has been reduced both in water and in iso-propanol solvents. As a result of silylation of NSFPA, the average dp in aqueous suspension increased from 100 nm to 143 and 124 nm for samples 3 and 4, respectively. Obviously, some aggregation between silanised particles occurred because of hydrophobic interactions between silylated nanoparticles in polar aqueous suspension. However, the reduction in average dp to 88.6 nm for sample 3 was achieved by ultrasonication of this suspension for 1 min. A similar effect was observed for samples 2 and 5 in isopropanol. For example, average diameter of aggregates for sample 5 dropped from 180.5 to 78.5 nm after sonication in isopropanol. A more dramatic increase in average dp from 50 nm to 200 nm was reported by Krueger et al. for DND aggregates treated with (3aminopropyl)trimethoxysilane [18]. Inter-molecule condensations can take place when using trimethoxysilanes for the modification of DND surface, leading to the formation of siloxane bonds between DND nanoparticles and, ultimately, to the formation of bigger aggregates. However, such condensation reactions are not possible when trialkylmonochlorosilanes or HMDS are used as silylating reagents. Interestingly, suppression of the zeta potential on the DND samples was observed following silylation. The aqueous suspension of NSFPA had a zeta potential of 29.6 mV, while the silylated samples 3 and 4 displayed lower values of 3.1 and 8.2 mV, respectively. 3.6. Capillary zone electrophoresis

Fig. 6. Suspension stability of DND type NSFPA, YTM and silylated (Sample 3) and (Sample 4) in dichloromethane: before sonication (above), an hour after sonication (middle) and a week after sonication (below, with shown Tyndall effect).

which may be attributed to the different binding environments due to the surface functionalisation reaction (Eqs. (1)–(4)). It should be noted that a lower Si concentration was recorded by XPS for the sample 4 as compared with data obtained by ICP-MS (Table 1). Obviously, this indicates a relatively poor stability of silylated DND samples during prolonged storage as suspensions in isopropanol as XPS spectra were obtained almost two years after ICP-MS analysis. 3.5. Suspension stability Silylated DNDs form stable suspensions after sonication in nonpolar solvents like dichloromethane. The suspensions remained stable for up to one week, and demonstrated effective light scattering due to Tyndall effect (Fig. 6), thus confirming the suspension stability. This behaviour differs from that observed for unmodified DNDs, which do not usually form stable suspensions (e.g. YTM type DND shown in Fig. 6 and results presented in [31]) and flocculate in less than an hour in dichloromethane. Particle size (dp ) measurements (by dynamic light scattering) (DLS) of the samples under investigation in both aqueous suspension and iso-propanol are presented in Table 2. There is a clear reduction in the average dp when YTM type DND is subjected to further purification procedures to produce the NSFPA material. The

CZE has recently been employed to characterise the dispersion quality and stability of DND, and was previously applied to the precursor NSFPA material [27]. In this earlier study, the NSFPA sample in sodium tetraborate buffer (pH 9.3) exhibited a well-defined broad peak followed by two smaller broad peaks, demonstrating the presence of stable agglutinates within a polydisperse sample (Fig. 7; 20 s injection of NSFPA at 5 kPa; +15 kV applied voltage; effective capillary length 57.5 cm). Following silylation of NSFPA, the samples were studied under similar conditions by CZE (5 s injection at 5 kPa, +5 kV applied voltage, effective capillary length 24 cm). Fig. 7 shows the typical electropherograms obtained for samples 3 and 4 in 20 mM sodium tetraborate, where the DNDs were negatively charged at pH 9.3, and migrated after the electroosmotic flow (EOF) (iso-propanol marker) in a single well-defined broad peak. The absence of any spikes within this broad peak demonstrates that the suspension is stable under these conditions, and that it is likely to be composed of small agglutinates as similarly observed for commercial DND samples in sodium tetraborate buffer [27]. This observation is in agreement with the particle size data obtained by DLS for these silylated DNDs in aqueous suspension (Table 2) as discussed above. Both silylated DND materials demonstrated a similar apparent mobility under the same conditions (calculated by sample − EOF where  = Ld ·Lt /V·tmigration wherein Ld is capillary length to detection window, Lt is total capillary length, V is applied voltage and tmigration is the migration time of the analyte). The apparent mobility of sample 3 and sample 4 were −2.58 × 10−8 m2 /Vs and −2.43 × 10−8 m2 /Vs, respectively. These similar mobility values indicate the similarity in both size and charge of the silylated DND under these conditions.

Table 2 Average particle sizes (dp , nm) of original and silylated DND measured by dynamic light scattering. Sample Silylating agent

YTM –

NSFPA –

(5) TMCS

(2) DMODCS

(4) DMODCS + HMDS

(3) TMCS + HMDS

Aqueous suspension Isopropanol suspension

264.3 171.8a

100.0 83.7a

– 180.5a (78.5)b

– 115.9a

143.1 –

124.3 (88.6)a –

a b

After ultra sonication for 1 min. After ultra sonication for 3 min.

404

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

Fig. 7. Electropherograms showing the separation of purified NSFPA sample silylated nanodiamond Sample 3 (TMCS + HMDS) and Sample 4 (DMODCS + HMDS) in 20 mM sodium tetraborate buffer pH 9.3. Conditions: 5 s injection at 5 kPa, +5 kV applied voltage.

3.7. Thermal stability The impact of vacuum annealing on the Si-content of the silylated DND was investigated. After careful washings with dry toluene and then acetone, three types of silylated DNDs were annealed under vacuum: namely those from procedures (3), (4) and (5), at 250 ◦ C for 18 h. After this continuous exposure, residues were deposited on the walls of sample tubes containing silylated DND from procedures (3) and (4) demonstrating that there was adsorption of HMDS on the surface of DND; whereas no residue was deposited onto the walls of sample tube (5) when only TMCS was used for silylation (Fig. 8). These results are in agreement with FTIR data showing an extra band at 1435 cm−1 for samples (3) and (4). Following annealing, there was a reduction in sample mass, and a notable reduction in Si-content was also observed, as determined by ICP-MS (Table 3). It should be noted that the temperature of 250 ◦ C used here is below the decomposition thermal limits of the correspondent silylation products, which is 300–350 ◦ C for TMCS-derivatives [41]. Additionally, the effect of acid cleavage (excess 0.5 M HCl in isopropanol overnight) on the annealed silylated DND was found to be a relatively slow, possibly, due to steric hindrance effects resulting from the pseudo-porous nature of the nanodiamond agglutinates (Table 3). TGA can offer further insight into the effect of high temperatures on the degradation of silylated DND. Fig. 9 shows the multi-step decomposition resulting from annealing of silylated DND (3) and (4) in nitrogen atmosphere. Following annealing to 900 ◦ C, sample (3) underwent ∼18% weight loss and sample 4 underwent ∼23% weight loss, thus demonstrating the DMODCS/HMDS silylated ND had more physically adsorbed HMDS molecules, which are generally less thermally stable under these conditions, than TMCS/HMDS silylated DND. The initial rapid loss seen for (3) can be attributed to

Fig. 8. Sample tubes with heated silylated DND samples (3), (4) and (5).

Table 3 Silicon content variation in silylated DND measured by ICP-MS, ␮g g−1 . Si in dry samples, ␮g g−1 a

Samples (silylating reagents)

Sample 3 (TMCS + HMDS) Silylated DND Silylated and heated at 250 ◦ C DND Acid washed a

Sample 4 (DMODCS/HMDS)

29,300 15,600

27,400 14,000

7500

5200

Sample 5 (TMCS) 7020 3950

1060 −1

Si-content in original NSFPA type of DND was 48 ␮g g

.

residual solvent or volatiles (loss to temperatures of 100 ◦ C). There is a slight gain in weight seen for (4) within this temperature range (up to ∼140 ◦ C). The two materials display different thermal behaviour from 200 to 500 ◦ C, where surprisingly (3) remains relatively stable with no significant changes in weight, and (4) undergoes a steady decline in weight. The favourable thermal properties of ND may play a role in stabilising (3), as the weight loss expected around 350 ◦ C for TMCS degradation was not evident in the TGA curve. ND is known to have improved heat dissipation in composite systems and it has excellent thermal conductivity and low thermal expansion [17]. As seen in the TGA curve, sample (3) was no longer stable above 500 ◦ C, and it experienced continual weight loss throughout the rest of the analysis. It is important to note that ND can undergo surface graphitisation in the temperature range 751–800 ◦ C in an inert atmosphere, and above this temperature range graphitisation of ND continues to produce concentric curved graphitic sheets, known as carbon onions [42]. This conversion may result in weight losses which can account for the losses observed in Fig. 9 above 750 ◦ C.

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

Fig. 9. Thermogravimetric curve for silylated DND samples (3) and (4) heated to 900 ◦ C in an atmosphere of nitrogen (50 mL min−1 ).

4. Conclusions In this work the Si concentration profiles (along with other common non-carbon elements) of seven differently purified, refined and silylated DND samples were determined and compared using a direct ICP-MS approach. The efficiency of different silylation procedures, preliminary preparations and post-silylation treatments were assessed with respect to the changes in covalently bonded and adsorbed silicon, through ICP-MS analysis and supporting methods such as FTIR, XPS, TGA and capillary electrophoresis. The ICP-MS method provided a fast and accurate non-carbon estimation, as well as an estimate of contamination levels and types. Determined elemental variations were considered with the purpose of improving procedures for DND purification, storage and functionalisation, which are essential steps in applying this unique nano-material in biomedicine, nano-composites and other fields. Acknowledgements This work was supported by grants from the Australian Research Council to ACROSS (DP110102046 and DP150101518) and CSL (LE0989539). The authors would also like to acknowledge the Central Science Laboratory (University of Tasmania) and Mark Wainwright Analytical Centre (University of New South Wales) for substantial instrumental support and service. Dr. Bill Gong (University of New South Wales) is acknowledged for his assistance with XPS spectra. References [1] O. Shenderova, D. Gruen, Ultrananocrystalline Diamond: Synthesis, Properties, and Applications, Elsevier, 2012, pp. 133–164. [2] O. Shenderova, D. Gruen (Eds.), Ultrananocrystalline Diamond, William-Andrew, Norwich, NY, 2006. [3] G.V. Sakovich, A.S. Zharkov, E.A. Petrov, Nanotechnol. Russ. 8 (2013) 581–591. [4] D.H. Wang, L.-S. Tan, H. Huang, L. Dai, E. Osawa, In-situ nanocomposite synthesis: arylcarbonylation and grafting of primary diamond nanoparticles with a poly(ether-ketone) in polyphosphoric acid, Macromolecules 42 (2009) 114–124. [5] O. Shenderova, C. Jones, V. Borjanovic, S. Hens, G. Cunningham, S. Moseenkov, V. Kuznetsov, G. McGuire, Detonation nanodiamond and onion-like carbon: applications in composites, Phys. Status Solidi (a) 205 (2008) 2245–2251. [6] E. Duffy, X. He, E.P. Nesterenko, A. Dey, S. Krishnamurthy, D. Brabazon, P.N. Nesterenko, B. Paull, Thermally controlled growth of carbon onions within porous graphitic carbon-detonation nanodiamond monolithic composites, RSC Adv. 5 (2015) 22906–22915.

405

[7] H. Mazeheri, S.R. Allahkaram, Deposition, characterization and electrochemical evaluation of Ni-P-nano diamond composite coatings, Appl. Surf. Sci. 10 (2012) 4574–4580. [8] A.A. Peristyy, O.N. Fedyanina, B. Paull, P.N. Nesterenko, Diamond based adsorbents and their application in chromatography, J. Chromatogr. A 1357 (2014) 68–86. [9] B. Kharisov, O. Kharissova, L. Chavez-Guerrero, Synthesis techniques, properties, and applications of nanodiamonds, Synth. React. Inorg. Met. Org. Nano Met. Chem. 40 (2010) 84–101. [10] V. Mochalin, O. Shenderova, D. Ho, Y. Gogotsi, The properties and applications of nanodiamonds, Nat. Nanotechnol. 7 (2012) 11–23. [11] H.B. Man, D. Ho, Diamond as a nanomedical agent for versatile applications in drug delivery, Phys. Status Solidi A 209 (2012) 1609–1618. [12] A. Krueger, Y. Liang, G. Jarre, J. Stegk, Surface functionalisation of detonation nanodiamond suitable for biological applications, J. Mater. Chem. 16 (2006) 2322–2328. [13] V. Dolmatov, Polymer-diamond composites based on detonation nanodiamonds, Part 2, J. Superhard. Mater. 29 (2007) 65–75. [14] L. Pramatarova, E. Radeva, E. Pecheva, T. Hikov, N. Krasteva, R. Dimitrova, D. Mitev, P. Montgomery, R. Sammons, The advantages of polymer composites with detonation nanodiamond particles for medical applications, in: L.D. Pramatarove (Ed.), On Biomimetics, Intech, 2011, pp. 297–320. [15] J. Lu, Y.H. Wang, J.B. Zang, Y.N. Li, Protective silicon coating for nanodiamonds using atomic layer deposition, Appl. Surf. Sci. 253 (2007) 3485–3488. [16] P.N. Nesterenko, O.N. Fedyanina, Y.V. Volgin, Microdispersed sintered nanodiamonds as a new stationary phase for high-performance liquid chromatography, Analyst 132 (2007) 403–405. [17] P.N. Nesterenko, P.R. Haddad, Diamond-related materials as potential new media in separation science, Anal. Bioanal. Chem. 396 (2010) 205–211. [18] A. Krueger, J. Stegk, Y. Liang, L. Lu, G. Jarre, Biotinylated nanodiamond: simple and efficient functionalization of detonation nanodiamond, Langmuir 24 (2008) 4200–4204. [19] Y. Liang, M. Ozawa, A. Krueger, A general procedure to functionalize agglomerating nanoparticles demonstrated on nanodiamond, ACS Nano 3 (2009) 2288–2296. [20] M.-G. Jeong, Y.-S. Chun, D.-S. Lim, J.Y. Kim, Effect of a silane coupling agent on the optical and the mechanical characteristics of nanodiamond/acrylic resin composites, J. Korean Phys. Soc. 65 (2014) 1049–1053. [21] V.Y. Dolmatov, Composition materials based on elastomer and polymer matrices filled with nanodiamonds of detonation synthesis, Nanotechnol. Russ. 4 (2009) 556–575. [22] D.P. Mitev, A.T. Townsend, B. Paull, P.N. Nesterenko, Direct sector field ICP-MS determination of metal impurities in detonation nanodiamond, Carbon 60 (2013) 326–334. [23] D.P. Mitev, A.T. Townsend, B. Paull, P.N. Nesterenko, Screening of elemental impurities in commercial detonation nanodiamond using sector field inductively coupled plasma-mass spectrometry, J. Mater. Sci. 49 (2014) 3573–3591. [24] H. Gaspar, C. Pereira, S. Rebelo, M. Pereira, J.L. Figueiredo, C. Freire, Understanding the silylation reaction of multi-walled carbon nanotubes, Carbon 49 (2011) 3441–3453. [25] M.V. Cagnoli, S.G. Casuscelli, A.M. Alvarez, J.F. Bengoa, N.G. Gallegos, M.E. Crivello, E.R. Herrero, S.G. Marchetti, Ti-MCM-41 silylation: development of a simple methodology for its estimation: silylation effect on the activity and selectivity in the limonene oxidation with H2 O2 , Catal. Today 107–108 (2005) 397–403. [26] G. Brieva, J. Campos-Martin, M. de Frutos, J. Fierro, Some insights on the negative effect played by silylation of functionalized commercial silica in the direct synthesis of hydrogen peroxide, Catal. Today 158 (2010) 97–102. [27] E. Duffy, D.P. Mitev, P.N. Nesterenko, A. Kazarian, B. Paull, Separation and characterisation of detonation nanodiamond by capillary zone electrophoresis, Electrophoresis 35 (2014) 1864–1872. [28] D.P. Mitev, A.T. Townsend, B. Paull, P.N. Nesterenko, Microwave-assisted purification of detonation nanodiamond, Diam. Relat. Mater. 48 (2014) 37–46. [29] S.S. Batsanov, E.V. Lesnikov, D.A. Dan’kin, D.M. Balakhanov, Water shells of diamond nanoparticles in colloidal solutions, Appl. Phys. Lett. 104 (2014) 133105. [30] C. Sweeley, R. Bentley, M. Makita, W. Wells, Gas–liquid chromatography of trimethylsilyl derivatives of sugars and related substances, J. Am. Chem. Soc. 85 (1963) 2497–2507. [31] A. Krueger, T. Boedeker, Deagglomeration and functionalisation of detonation nanodiamond with long alkyl chains, Diam. Relat. Mater. 17 (2008) 1367–1370. [32] S. Fruehauf, I. Streiter, S.E. Schulz, E. Brendler, C. Himcinschi, M. Friedrich, T. Gessner, D.R.T. Zahn, Advanced metallization conference, in: A. McKerrow, Y. Shacham-Diamand, S. Zaima, T. Ohba (Eds.), Mater. Res. Soc. Conf. Proc. ULSI XVII, MRS, Warrendale, PA, 2011, pp. 287–294. [33] P.G. Harrison, A. Torr, FT-IR study of the gas phase thermolysis of trimethylsilyl acetate, J. Organomet. Chem. 538 (1997) 19–29. [34] A.P. Nascimento Filho, A.T. de Carvalho, M.L.P. da Silva, N.R. Demarquette, Preconcentration in gas or liquid phases using adsorbent thin films, Mater. Res. 9 (2006) 33–40. [35] T. Oh, K. Lee, C. Choi, A study on the structure of Si O C thin films with nano size pore by ICP CVD, Stud. Surf. Sci. Catal. 146 (2003) 387–390.

406

E. Duffy et al. / Applied Surface Science 357 (2015) 397–406

[36] J.A. Zhang, D.S. Su, R. Blume, R. Schlogl, R. Wang, X.G. Yang, A. Gajovic, Surface chemistry and catalytic reactivity of a nanodiamond in the steam-free dehydrogenation of ethylbenzene, Angew. Chem. Int. Ed. 49 (2010) 8640–8644. [37] T. Meinhardt, D. Lang, H. Dill, A. Krueger, Pushing the functionality of diamond nanoparticles to new horizons: orthogonally functionalized nanodiamond using click chemistry, Adv. Funct. Mater. 21 (2011) 494–500. [38] T. Petit, J.C. Arnault, H.A. Girard, M. Sennour, P. Bergonzo, Early stages of surface graphitization on nanodiamond probed by X-ray photoelectron spectroscopy, Phys. Rev. B 84 (2011) 233407.

[39] S. Tanuma, C.J. Powell, D.R. Penn, Calculations of electron inelastic mean free paths. VIII. Data for 15 elemental solids over the 50–2000 eV range, Surf. Interface Anal. 37 (2005) 1–14. [40] Yu.V. Butenko, S. Krishnamurthy, A.K. Chakraborty, V.L. Kuznestsov, V.R. Dhanak, M.R.C. Hunt, L. Siller, Photoemission study of onion-like carbons produced by annealing nanodiamonds, Phys. Rev. B 71 (2005) 075420. [41] GC Derivatization, Applications Handbook & Catalog, Pierce Biotechnology, Inc., Rockford, IL, 2003. [42] N.S. Xu, J. Chen, S.Z. Deng, Effect of heat treatment on the properties of nano-diamond under oxygen and argon ambient, Diam. Relat. Mater. 11 (2002) 249–256.